{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

B_StaticLSEsts&amp;Heckit

# B_StaticLSEsts&amp;Heckit - Cross-Section Regression...

This preview shows pages 1–2. Sign up to view the full content.

Cross-Section Regression Estimates of Labor Supply Elasticities: Procedures and Problems Beginning in the 1970’s labor economists tried to learn about how households’ labor supply decisions respond to offered wages by running cross-section regressions in survey samples of (usually) thousands of individuals. The idea was that different people in a population face different prices ( w ) for their labor and are endowed with different amounts of non-labor income ( G ). By asking whether people who earn higher hourly wages work more hours, economists hoped to learn whether a “representative” household would respond to a permanently higher wage by working more or less (i.e. to estimate the sign and magnitude of the uncompensated labor supply elasticity). This exercise generally had two main goals (the first of which was probably most prevalent): 1. Estimating a parameter (the size of the labor supply response to wages) that is useful to know for policy purposes, whether or not any particular theory of labor supply behavior –including the classic, static model— is correct. For example, labor supply elasticities tell us how much (or whether) tax cuts will stimulate extra work effort, and help us estimate who “really” bears the burden of payroll taxes. 2. Testing the predictions of the static labor supply model. Some practitioners (e.g. Ashenfelter and Heckman) used the estimated responses to G and w in the data to compute a pure substitution effect, and checked to see if it had the predicted sign. Before going into the details, it is worth reflecting on the notion of estimating a price elasticity (in this case the price is the wage) from pure cross-section data. If the price in question were, say, the price of home heating oil and the sample was all the households in a city, this would be a pretty hopeless exercise. The reason is that, at any given point in time, competition between firms tends to ensure that all households pay the same price. There would be no price variation in your sample from which you could estimate how households behave when they face different prices. In the labor supply case, different persons do face different prices for their labor in a cross-section of individuals, for various reasons such as different family backgrounds, intelligence, motivation, and past education decisions. So we can ask whether, at any point in time, high-wage people work more or fewer hours. But you should also immediately be suspect of this idea: high-wage people may work more hours than low-wage people because they differ from low-wage people in many ways other than just their wage, and controlling for all these other differences may not be that easy. So it is unclear we can really isolate a causal effect of wages on labor supply this way. We will come back to this problem under the heading of “omitted variables bias” below. First, let’s describe what a typical economist did.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}

### Page1 / 11

B_StaticLSEsts&amp;Heckit - Cross-Section Regression...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online